they were essentially looking at the small signal model as VS describes it in his paper. But the moreso appropiate model (talking power amps here which is what this paper addresses) for the primary L is that... in a large signal model the primary L is in parallel with the reflected impedance of the output transformers...

I spent much time making these points about the need for L in a large signal model and how to moreso accurately model or describe those interactions and why having a sufficiency of primary inductance was critical... I promise not to ever do those monster threads again...

but... what I would like to point out... is that the guinevere is perhaps more appropiately used in a small scale model... and does NOT employ an output transformer...

from a small signal model (where you are not delivering power) then the plate choke can be modeled as being in parallel with the r sub p of the preamp tube...

If the r sub p of the tube is say 2300 ohms... then a plate choke with about 180 henries will have an AC impedance at twenty hertz of approx ten times the magnitude of the r sub p of the tube... and NOT load the tube in any significant manner.

make a long story shorter... plate chokes have been widely used with success... I won't argue the technicals of this with anyone...

is it worth a try? You bet. Is it simple to try... mechanically yes... expense wise... no. The CCS will be far less expensive to implement than buying a pair of appropiate plate chokes. But plate chokes are an option and perhaps an attractive SONIC option vis-a-vis a CCS load.

cheers,

msl

And AC choke (with as little as say 500H all the way up to 5000H) in parallel with the output capacitor would enable you to reduce the size of the 4.7mfd output coupling capacitor.

with the 4.7mfd cap and a 500H ac choke in parallel.... the resonant frequency would be 3.28 hertz.

so...say you back off of the cap size and reduce it to 1mfd. then with a 500H choke your resonanct freq would be 7.11 hertz. More than a full octave lower than twenty hertz if you take that to be your lowest freq of interest.

a 1mfd cap with a 1000 henry choke would get your resonance down to 5.03 hertz.

a .47mfd cap with a 1000 henry choke would get your resonance down to 7.34 hertz.

increasing the choke's L while keeping the c constant lowers the resonance. the lower the amount of L the larger the C needed for the same resonance point.

but the beauty here may be that you could use caps as small as say .22mfd or .47mfd with an appropiately sized ac choke (in the output part of the circuit) and still get response well below the audio band.

this may yield sonically even greater improvements than just switching to a plate choke as opposed to using the CCS as drawn.

MSL

>

So then, what is the point where the transition from the unloaded small signal model gets swapped for the power/large signal one? Specifics, and some numeric analysis would be quite useful if you can.

What does the transition from small to large signal look like and how is it best described? some detail please. The load is always in parallel with the plate resistance. At what power point does the load start changing the response?

This point is valid for an ideal inductance. the self resonance puts limits on deliverable impedance. it is not omega*L but is it going to be enough to consider it as that?

***but... what I would like to point out... is that the guinevere is perhaps more appropiately used in a small scale model... and does NOT employ an output transformer...****

So then, what is the point where the transition from the unloaded small signal model gets swapped for the power/large signal one? Specifics, and some numeric analysis would be quite useful if you can.

***from a small signal model (where you are not delivering power) then the plate choke can be modeled as being in parallel with the r sub p of the preamp tube...***

What does the transition from small to large signal look like and how is it best described? some detail please. The load is always in parallel with the plate resistance. At what power point does the load start changing the response?

****If the r sub p of the tube is say 2300 ohms... then a plate choke with about 180 henries will have an AC impedance at twenty hertz of approx ten times the magnitude of the r sub p of the tube... and NOT load the tube in any significant manner.***

This point is valid for an ideal inductance. the self resonance puts limits on deliverable impedance. it is not omega*L but is it going to be enough to consider it as that?

***make a long story shorter... plate chokes have been widely used with success... I won't argue the technicals of this with anyone...***

What technicals are those? That people have built amps that sound good and happen to have used plate chokes is not a disputed item. Relax...there is no fight here.

***is it worth a try? You bet. Is it simple to try... mechanically yes... expense wise... no. The CCS will be far less expensive to implement than buying a pair of appropiate plate chokes. But plate chokes are an option and perhaps an attractive SONIC option vis-a-vis a CCS load.***

Interesting things are always worth trying. short story: know as much about what you are experimenting with in order to get meaningful results.

cheers,
Douglas

The 1meg resistor is not doing much of anything save providing a means of ref'ing ground to the output. The bulk of this should be done at the amp. Here a grid choke is a good spot to use Iron. They're easy to wind and not too big. No DC, small AC...easy choke to design and execute in Cu and Fe.

Before throwing in a large value grid choke, one should be aware of the reasons for emloying grid inductances in the first place: grid current! in small doses, but grid current. It is also AC in nature, so the big value grid chokes will not do much better than a resistor in this arena.

Curiously, if a low value resistor can be driven by the source, it is likely to be a better option. Tube sources capable of driving few kOhm loads are few and far between. Also to be considered is the effect of exerting this 'effort' to drive a low-numeric resistive load. Does the inductor allow better overall performance? Depends on what it is being driven by, and of course personal preference.

There is no arguement over the choices made under personal taste. this idea seems foreign sometimes, but it is a good thing to keep in the back of one's mind whilst discussing 'improvements' and 'better'...

Like anything else, design for what is required. When what is required is shrouded in mystery and/or religous fanaticism, good experimental method and analysis from the first principles is required. If you want it done right, do it yourself. Just educate yourself adequately is the unspoken requirement.
cheers,
Douglas

So Mike, you deliver this interesting idea for a new linestage circuit, and then can't deliver a well thought out reason past, "plate chokes sound different to some", and "plate chokes are simpler than CCS-es".

Keep up the good work if you please, I would like to know what sort of plate choke can deliver the goods. Keep in mind, the Guinevere circuit delivers some very nice numbers to back up her fine sonics.

In God we trust.
Everybody else: bring data.

cheers,
Douglas

What the choke offers is far different from what the CCS offers in a few areas. Not that either is better...just better *WHY* is what I was after. But I didn't really expect an answer. I don't think he has one, let alone one he'll share.

There is obviously room to play with the Guinevere power supply. Hell, it makes some amps look positively weak...
cheers,
Douglas